GENDER CLASSIFICATION FROM FACE IMAGES

Authors : Eyyüp Yildiz, Tolga Ensari

Pages : 31-42

View : 18 | Download : 8

Publication Date : 2017-04-28

Article Type : Other

Abstract :In this article, we study on gender classification which is one of the important issue in security, statistics and related commercial areas. In the study, FEI face data set has been used that has 200 female and 200 male frontal face images. Principal component analysis (PCA) has been used for feature extraction process. We use all part of the face images instead of taking some part of them. Support Vector Machine (SVM) and k-nearest neighbor algorithms used for classification test phases. We compare the results which obtained in our experiments and give them in tables and graphs. According to the experiments, defined as hybrid method principal component analysis with k-nearest neighbor method gives better recognition accuracy then defined as hybrid method principal component analysis with support vector machine method.
Keywords : Gender classification, face recognition, principal component analysis, k-nearest neighbor